CUDA/summary_threads.cu at master · 4ubi/CUDA · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
#include <cuda.h>
#include <stdio.h>

#define CUDA_CHECK_RETURN(value) {\
        cudaError_t _m_cudaStat = value;\
        if (_m_cudaStat != cudaSuccess) {\
        fprintf(stderr, "Error %s at line %d in file %s\n",\
        cudaGetErrorString(_m_cudaStat), __LINE__, __FILE__);\
        exit(1);\
}}

__global__ void gTest(float* a){
        a[threadIdx.x+blockDim.x*blockIdx.x]=(float)(threadIdx.x+blockDim.x*blockIdx.x);
}

__global__ void sum (float* a, float* b)
{
a[threadIdx.x+blockDim.x*blockIdx.x]+=b[threadIdx.x+blockDim.x*blockIdx.x];
}

int main(int argc, char* argv[]){
        float *da, *db, *ha;
        int num_of_blocks=1<<2, threads_per_block=1<<2;
        int N=num_of_blocks*threads_per_block;
        float elapsedTime;

        cudaEvent_t start,stop;
        cudaEventCreate(&start);
        cudaEventCreate(&stop);

        ha=(float*)calloc(N, sizeof(float));

        CUDA_CHECK_RETURN(cudaMalloc((void**)&da,N*sizeof(float)));
        CUDA_CHECK_RETURN(cudaMalloc((void**)&db,N*sizeof(float)));

        gTest<<<dim3(num_of_blocks), dim3(threads_per_block)>>>(da);
        gTest<<<dim3(num_of_blocks), dim3(threads_per_block)>>>(db);

        cudaEventRecord(start,0);
        sum<<<dim3(num_of_blocks), dim3(threads_per_block)>>>(da,db);
        cudaEventRecord(stop,0);
        cudaEventSynchronize(stop);
        CUDA_CHECK_RETURN(cudaGetLastError());

        cudaEventElapsedTime(&elapsedTime,start,stop);
        fprintf(stderr,"gTest took %g\n", elapsedTime);
        cudaEventDestroy(start);
        cudaEventDestroy(stop);

        CUDA_CHECK_RETURN(cudaMemcpy(ha,da,N*sizeof(float), cudaMemcpyDeviceToHost));
        for(int i=0;i<N;i++)
                printf("%g\n",ha[i]);
        free(ha);
	cudaFree(da);
	cudaFree(db);
        return 0;
}